M17x Refresh Discussion

Yeah, AMD had confirmed it a while ago, but the first quarter goes all the way to march, and now I believe there's a chance that it will be available from beginning of january, which makes it extra cool =D

 

that means mxm 2.1 and mxm3.0 are the same i may put some 4870's in my m17-r1

 

A Mobility 5800 is not in the M17x yet.

Even if Fuad's speculation is right about Alienware it could be the M15x that gets one first as an ATI option or even a new M21x model.

Then you've got the traditional delay that comes between almost every notebook's launch and their sale. e.g. 90 day countdown to M17x launch; 3 months after that for Crossfire option; M15x; Adamo(XPS); Clevo M980nu(shown at CeBit Mar 2009 for sale July 2009); etc.


Then you can throw ~50% yield rates for 40nm at TSMC on top of all that.

 

That's true, but I'm really hoping that it is alienware, it is a crossfire notebook, and it won't be that much more expensive over current price, and that it goes for sale as soon as January =D

Then again, we all can dream, right?

 

I don't believe delivery of 58XX would come in July. You need to understand that Dell is more reactive than Alienware and the competition is already starting to take orders for laptops with 5850's cards.

I say order around Christmas for delivery in January or early February.

Remember that Dell has upgraded their XPS line of laptops in the past every 6 to 8 months in terms of graphics card option / CPU.

The design may stay the same and it might not be a refresh but definitely we will see 58XX in the next 3 months.

 

Two things: one, is ATI a subdivision of AMD? And how long has this been the situation? I must be going senile or something to miss that.

 

I look forward to the M17x refresh, whenever that is. Hopefully not to far in the distant future though.

 

hmm... AMD bought ATI a long time ago... I think it was back when ATI was releasing their ATI Radeon 3XXX series, or maybe even before that!

 

Roger that. That would explain all the AMD/ATI motherboard proprietary compatibilities I've been seeing.

 

Forget the duals and quads - let the fun begin :
http://www.tomshardware.com/news/Intel-Cloud-Cores-Processor-CPU,9193.html

 

All I know is the step up to four cores didn't do a thing for games optimized for two. In fact, I have publications that stated their test revealed a slight decrease in performance.

 

48 core processor, 45nm size, 10-20 times the power of current intel cores, with the power consumption of two light bulbs?

So..why aren't we using this right now again? Screw i7 xD

 

Well, price isn't listed. If each chip costs, oh, half a million, that's probably why. XD

 

Most likely isn't x86 architecture so not compatible with any current os. Possibly problems with chip or motherboard. The lack of x86 would partially explain why it takes so little power.

they should of named it The Brain.

 

Anyway, the refresh is going to happen sooner than most of us thought,...given http://www.fudzilla.com/content/view/16684/41/

 

Actually, according to Gizmodo it can run x86 software:

"Dubbed as the "Single-chip Cloud Computer" (SCC), the 1.3-billion transistor processor one ups it's 80-core Polaris predecessor because it can run standard x86 software. So far, it has successfully booted Windows and Linux during demonstrations."

Even if it didn't, this seams to be a big leap in terms of processor computing, and OS support would be insignificant to debate as new iterations of processors always get supported.

My question is...why are they working on an i9, and seemingly not considering this for mass reproduction.

 

i don't think i want an 18" screen plus i like the backlit keyboard of m17x

 

The answer is simple: money.
The strategy is to always stay just a little ahead, every 5% worth billions, - why skipping them? Unless, of course, there is a risk that your competitors will be there first.
Actually, a friend of mine @Intel R&D was writing algorithms for 64-128 core CPUs back in 2006.
The problem is that both Intel and AMD are too close to the limit in silicon tech. And switching entirely to a new tech. (biochips+neurons, fiber-optics etc.) will require some doozies. So they will use the good old rebranding tactics and increase the number of cores/threads till the new tech. is ready.
Another question - Why do they show something so advanced now? I'm sure those are just simple marketing games. But who knows...

 

They're not releasing it now because of Moore's Law.

 

In 10 years we will have a gaming PC the size of a netbook except it won't have a screen but rather a holographic projection device enabling the user to interact in 3D with all the objects.

Basically to play FarCry 8 you will be be the player, you will feel the bullets going into you and you will need a real doctor next to you to revive you...

 

Since Moore's Law is going from linear to exponential growth, there is no objection to 9 to 48 cores jump.

 

No, that's not Moore's Law. Moore's Law is a conceptual/perception on the business side of things. I'm sure Intel has processors downstairs in a bunker somewhere that can power rockets through telepathy alone. What they show us is just enough to keep us interested and squeezing our wallets for that extra penny.

 

Never said what it is, I said that the tendency of its growth is exponential now.
You can find it even in your own link, if you read it.
But I'm not here to argue about the definition, especially since the second part of your statement is exactly what I think.
If you wanna discuss Moore's Law, we could open a side topic. But IMHO it is no longer the correct approach.

P.S.> Intel has neither telepathy operated CPUs nor any other full fledged technology. They do work in a few directions though...
I know this from a trustworthy source, from those bunkers you mentioned. In 5 years there will be a new tech. ready to be deployed but not yet.

 

Ditto.

By the way, do you wear teh AW hat often? Do people ask you what it is? And do you have to explain to them it is the current top-rated gaming computer for 2009 by several publications?

 

Lol, the hat is awesome, Alienware can now easily make millions from just selling these.
As for the rest, I always like to put my own photos on forums' avatars, but since I also needed to stay unidentified as much as possible, - so had to put on everything I had nearby at that moment.
Looks retardish, I know....Feels too
And by the way, here in Calgary it seems no one knows about computers in general (other than MACs) so I'm safe..

 

if i7 (not including arrandale) comes to m17x this refresh is it or isn't it going to have a igp. The m15x doesn't have an igp and if the m17x is the same way i don't want the i7 in the refresh. I know that the m15x stealth mode underclocks the cpu and gpu but i really think if the m15x had a igp and completely shot off the gpu it would have a substantially longer battery life. Does nvidia have an i7 chipset coming out?

 

Nvidia can't make chipsets for any Intel CPU newer than Monty.

Arrandale will have an IGP but it's on a 45nm die on the same PCB as the CPU so you only get it if you go for the dualcores. A quadcore i7 in the same machine wouldn't have the IGP.

 

where did the delete go?

 

Not that Nvidia chipsets were much to rave about anywho. IGP is definitely nice(if it were not for the heat), but I can do with underclocking the CPU/GPU for stealth.

 

The delete function has been removed on the AW forum

 

Without going to far off topic any clue as to why the function was removed ? i would have thought it was for the better myself ..

 

Actually Quad, its not just the AW forum . Its forum wide. Andrew posted on this in this thread. Senior members did have the ability to remove their posts if the post life was under 5 minutes. In addition to regular members (who never had the delete option), this has been removed for senior members.

 

FWIW, PC Mag seems to think that the m17x will be getting the i7 treatment "shortly"...

http://www.pcmag.com/article2/0,2817,2356595,00.asp

[Apologies if this was mentioned elsewhere in this forum.]

 

Yet another funny (pro M15X) review. I wouldn't rely on it.
They can only guess ...
"Demolished" in terms of CPU power is ...
Now, I have to ask you a few questions:
Is 5-20% improvement in CPU power can be considered "demolishing"?
Is 5-20% improvement in GPU power ..... "demolishing"?
Now for some stats.:
That CinebechR10(multi-rendering) of 14k+ was achieved with OC'ed (12k@stock) i7 920XM compared to STOCK i7 975 of 19K+ and STOCK QX9300 C2Q of 11K. Now ask our OC'ers on this forum to run their CineBench on [email protected] and compare. Same for 3DMark06/Vantage. A slight advantage in WPrime/SuperPi exists of course (otherwise there would be no point of even advertising the mobile i7 in enthusiast sector (other then slim-notebooks-segment), but I'll still sleep well knowing that in 1M SuperPi my P8700 gets 17 instead of 13 (920XM).
If Dell has a slightest strategic orientation, they will and must skip i7 refresh for M17X (the GPU refresh is still welcome) and focus on design and cooling improvements for the new platforms aka Fermi/Larabee.
If you buy M17X/M15X, you probably game and for that C2Q (even C2D) will be more than adequate for a few years.
i-Core MOBILE family is not a huge and important step forward, but in a couple of years from now there will be a new platform consisting of hundreds (weak) cores on a single chip capable of a huge performance leap (but of course we'll be getting the same minimalistic boost each year). Going back to the gaming - even GPU update is getting close to overkill in large notebooks. With 5870 we'll be able to run every game on 1900rez and full AA (probably even without CF).
Then what, guys? There's no update on XBOX/PS3 for at least couple of years; games are ported to and fro, - they will not require more than we already have...
2 cents...
I hope to prove wrong
But, today my good friend made up his mind for an upgrade of his desktop pc - to C2Q Q9650+single 5870 card. I did and he answered "I love the performance of i7 975 but I will hardly ever need anything better than C2Q and even Q9650 is overkill, but I like to have some room for OC'ing." And trust me that guy is THE GAMER.

 

Er, where do you find that that result was with an overclocked 920XM? The article specifically stated that you cannot overclock the M15x except with third party tools. Their result for the M17x QX9300 was also only 8309 in CinebenchR10. If you're pulling numbers from Notebookcheck, I have to say that I'm not comfortable with their 920XM numbers, mostly because they're based on a single benchmark. I mean, according to that single benchmark, the i7-720QM and i7-820QM are both faster than the 920XM in wPrime, with scores of 31/994 (32/1024) and 28/879, respectively, versus 40/1269 (!). Notebookcheck also says your P8700 should only get a 23 for SuperPi 1M.

 

And third party tools are never used for OC'ing?
I hope Mandrake will not see this...
11k for QX9300 is based on several benchmarks (all around 11k)
Their Wprime test was based on 1 or 2 so I don't rely on it.
P8700 gets 23 in SuperPi 1M 100%, but my CPU is OC'ed to 3.1Ghz and also gets 2900 in 3DMark06 instead of 2200
But you know, we can start a new thread : Extensive comparison of i7 920XM vs C2Q QX9300. what do you say?

 

Well, I just meant that if they had overclocked it, after mentioning that there was no way to do it without using a third party tool, I'm sure they would have mentioned that they did overclock it with X tool. The fact that they didn't implies to me that that 920XM wasn't overclocked, at least not with said third party tool. If by "overclock" you just meant "built in turbo-boost" then I won't disagree.

I don't mind notebookcheck's numbers on the QX9300, as I said, I just don't trust their 920XM numbers. 11K for a QX9300 is probably not too far off, I was just pointing out that for whatever reason, on their comparison chart, their QX9300 was only 8309.

An i7-920XM vs C2Q QX9300 thread could be fun, but I'm not sure I could convince myself to plop down the resources needed to actually have the both of them to compare, so we'd just be pulling things from other people, and the variables at that point could start getting messy.

But anyway, turning back somewhat to the subject of M17x refreshes, is there any reason your gamer friend didn't go with a Crossfire setup for his new machine? Could that be a hint that maybe a M17x refresh should move away from SLI and over to a single GPU solution?

 

I'm not so sure a MOBILE version of a 5870 would run everything on 1200p 4xAA by itself (no crossfire)... But then again, I am VERY willing to be proved wrong
I'm just waiting for CES 2010 to see if Alienware shows us a GPU refresh, since it's been confirmed that mobile ATI 58x0 series will be launched at it (January 7th!)

 

He didn't go for the CF option simply because it's not needed.
For 2 years from now (and in gaming world it's FOREVER) he will be able to run every game maxed out. And if there will be another Cr@zis, well you can always add another card.
In mobile sector we should probably hope and pray that our 5870x2 will be of equal power to a single desktop 5870.

By the way, we recently did bench-game-comparison between my 280M SLI vs his single desktop 4870. He won on every turn.

 

LOL. The performance gap that exists between Mobile GPUs and Desktop GPUs is ridiculously large. I guess we should be glad that, even with that enormous gap, mobile GPUs can be quite capable of running today's and tomorrow's games on high settings, high resolution and with good framerate (just make sure you have SLI/Crossfire on your notebook )

 

i agree with you

 

take the 9400 out and put an ati igp in so I can switch between discrete without having to restart computer while leaving c2d processor in it for ati graphics.

The biggest reason i haven't plopped down any money for the m17x is the screen in it. Last thing I want is a screen that isn't top of the line for how much I'm spending.
Rgb led screen not wled

Also what would you think if they refreshed the m17x with hydra chipset?

 

Does that exist on notebooks? I heard it barely exists on desktops themselves...

 

Actually those wled are pretty good after all. What would be good is the option to have anti-glare screens.

 

i don't know but if alienware contacted msi to make a custom board for the m17x then yes all they need is to put the lucid chip on the m17x mainboard. Though it would be on a core-i7.

wled screens would be nice but i hear rgbled have better colors. if you add an anti-glare coating to the mix the colors of any screen are going to be quite a bit duller. So i think it should be anti-glare coating plus rgbled screen to counter the color wash out

 

I also used to have a 4870, and unless something changed it isn't the whitewash you make it out to be, in fact a pair of mobile 4850s (basically 4830s) or the GTX280m often come out a little faster (although obviously thats SLI/Crossfire dependant).
Yeah desktop vs laptop right now is very unbalanced, but thats because new hardware has come out, and we're still waiting for new laptop chips. When new laptop chips come out it'll get much closer again.
Typically high end laptops are somewhere between a generation and half a generation behind (the GTX280m is roughly a desktop 9800GTX)

Nvidia and rebranding haven't exactly helped matters really, if they'd released a GT200 mobile chip we'd probably be in a better position already, but with Nvidias insistance on rebranding rather than reworking we're in a bit of a sticky position. The GTX280m was originally going to be the GTX180m which would have been more representative but they changed it for better marketing potential.

Once we have some new laptop chips we should be looking at somewhere between a GTX280 desktop and 5870 desktop performance in the crossfire/SLI models, and that's certainly not that far behind the curve at all, considering the hardware many people run (midrange cards,sometimes a generation or two behind).

Whichever way you look at it, we're in a much stronger position with laptops than we used to be, and with continuing efforts at miniaturisation and reducing power consumption, thats a trend that can only continue.

 

Nvidia's 40nm mobile chips are G200 based....they're just nerfed G200 tech because the top end GPUs couldn't be shrunk.

 

I'm pretty sure all laptop screens today are glossy, I haven't seen a matte screen for a long while.

Even Apple removed matte as an option completely, and moved towards one standard. It'd be wonderful to have a screen with the crisp colors of a glossy, but with some new form of anti-glare techniques.

 

my latitude e5400 has a matte finish.

 

You know what I meant...a decent GT200 part/lineup, rather than those few low-end parts they released,nerfed to oblivion and ironically slower than the G92 ones.